Detecting visually significant cataract using retinal photograph-based deep learning

Age-related cataracts are the leading cause of visual impairment among older adults. Many significant cases remain undiagnosed or neglected in communities, due to limited availability or accessibility to cataract screening. In the present study, we report the development and validation of a retinal photograph-based, deep-learning algorithm for automated detection of visually significant cataracts, using more than 25,000 images from population-based studies. In the internal test set, the area under the receiver operating characteristic curve (AUROC) was 96.6%. External testing performed across three studies showed AUROCs of 91.6-96.5%. In a separate test set of 186 eyes, we further compared the algorithm's performance with 4 ophthalmologists' evaluations. The algorithm performed comparably, if not being slightly more superior (sensitivity of 93.3% versus 51.7-96.6% by ophthalmologists and specificity of 99.0% versus 90.7-97.9% by ophthalmologists). Our findings show the potential of a retinal photograph-based screening tool for visually significant cataracts among older adults, providing more appropriate referrals to tertiary eye centers.

Evaluating the associations between obesity and age-related cataract: a Mendelian randomization study

BACKGROUND

The obesity-cataract association has been inconsistently reported. The fat mass and obesity-related (FTO) single-nucleotide polymorphism (SNP) rs9939609 is a major SNP associated with obesity and has been used as an instrumental variable for obesity in a Mendelian randomization (MR) approach. An interaction between the FTO SNP and macronutrient intake for obesity was suggested previously.

OBJECTIVE

The aim of this study was to assess the associations between obesity and cataract, using FTO SNP rs9939609 as an instrumental variable in an MR approach, and explore interactions of this SNP with macronutrient intake in relation to risk of cataract in a population-based cohort.

METHODS

The Blue Mountains Eye Study (BMES) is a longitudinal population-based study of common eye disease. Of 3654 baseline participants of the BMES (1992-1994), 2334 (75.8% of survivors) and 1952 (76.7% of survivors) were followed 5 and 10 y later. During the 5-y follow-up, 1174 new participants were examined. Cumulative cataract was defined as the presence of cortical, nuclear, or posterior subcapsular (PSC) cataract at any visit, following the Wisconsin Cataract Grading System. Imputed dosage of the FTO SNP rs9939609 was used. Quintiles of macronutrient intake (carbohydrates, protein, fats) were derived from an FFQ. ORs and 95% CIs were estimated using multivariable-adjusted logistic regression models.

RESULTS

After multivariable adjustment, there were no associations between BMI and any cataract types in MR models using rs9939609 as an instrumental variable. However, an interaction between rs9939609 and protein intake for PSC cataract risk was suggested (P = 0.03). In analyses stratified by quintiles of protein intake, each minor allele of rs9939609 was associated with increased odds of PSC (OR: 2.14; 95% CI: 1.27, 3.60) in the lowest quintile subgroup only.

CONCLUSIONS

Obesity was not causally associated with age-related cataract. However, among persons in the lowest quintile of protein intake, obesity may be associated with PSC cataract.

The Combination of Healthy Diet and Healthy Body Weight Is Associated with Lower Risk of Nuclear Cataract in the Blue Mountains Eye Study

BACKGROUND

Greater adherence to dietary guidelines has previously been found to be associated with decreased risk of visual impairment. However, whether or not this association extends to age-related cataract, 1 of the leading causes of visual impairment, is unknown.

OBJECTIVES

The aim of this study was to assess the association between adherence to dietary guidelines, using total diet score, and incidence of age-related cataract.

METHODS

Of 3654 baseline participants of the population-based Blue Mountains Eye Study cohort (1992-1994), 2334 (75.8% survivors) and 1952 (76.7% survivors) were examined after 5 and 10 y, respectively. Cataract was assessed from lens photographs using the Wisconsin Cataract Grading System. Baseline total diet score was calculated from FFQ data following a modified version of the Healthy Eating Index for Australians. OR with 95% CI were estimated using discrete logistic regression analyses, adjusting for age, sex, and other confounders. To test interaction, a cross-product term of 2 factors was included in regression models.

RESULTS

Of 2173 participants (84.7% of those returned for 1 or both follow-ups) with total diet score estimated, 57% were women, mean baseline age was 63.9 ± 8.4y, and mean baseline BMI was 26.3 ± 4.3 kg/m2. After multivariable adjustment, baseline total diet score was not associated with incidence of any cataract. A multiplicative interaction was observed between total diet score and BMI for incident nuclear cataract (P-interaction = 0.04): increasing baseline total diet score was associated with decreased risk of nuclear cataract among participants with BMI <25 (per unit increased total diet score, OR: 0.90; 95% CI: 0.81, 0.99; P = 0.02), but not among participants with BMI ≥25 (OR: 1.00; 95% CI: 0.92, 1.10; P = 0.95).

CONCLUSIONS

Adherence to dietary guidelines had no appreciable influence on cataract development overall in this older Australian population. However, adherence to dietary guidelines combined with healthy BMI is associated with decreased risk of nuclear cataract, an aging marker.

Systemic medications and cortical cataract: the Singapore Epidemiology of Eye Diseases Study

BACKGROUND/AIMS

To evaluate the association between systemic medications and cortical cataract prevalence in an Asian population.

METHODS

The Singapore Epidemiology of Eye Diseases Study recruited 10 033 Chinese, Malay and Indian residents aged 40+ years living in Singapore. Information on medication use was collected at interview using questionnaires. The presence and severity of cortical cataract were assessed from lens photographs using the modified Wisconsin Cataract Grading System. Associations between medications and the presence of cortical cataract were assessed using logistic regression. Associations between medications and greater severity of cortical cataract (none, minimal, early and late) were assessed using ordinal logistic regression.

RESULTS

A total of 8965 participants were included, the mean age was 57.6 (SD=9.8) years, and 4555 (50.8%) were women. After adjusting for age, gender, ethnicity, body mass index, smoking status, socioeconomic status, hypertension, hyperlipidaemia, diabetes, duration of diabetes and cardiovascular disease, ACE inhibitors (OR=1.27; 95% CI 1.05 to 1.55), fibrates (OR=1.57; 95% CI 1.05 to 2.35), alpha-glucosidase inhibitors (AGIs) (OR=1.85; 95% CI 1.13 to 3.02) and insulin (OR=1.80; 95% CI 1.11 to 2.93) were significantly associated with the presence of cortical cataract. Further adjusting for concurrent medication use did not alter these associations. Consistently, the four medications were also associated with a greater severity level of cortical cataract.

CONCLUSION

ACE inhibitors, fibrates and AGIs were associated with increased prevalence of cortical cataract in this Asian population, independent of the presence of hypertension, hyperlipidaemia and diabetes, respectively. Whether they contribute to the risk of cortical cataract needs confirmation in longitudinal studies.

Aldose Reductase Polymorphisms, Fasting Blood Glucose, and Age-Related Cortical Cataract

Purpose

To determine whether there is an association between polymorphisms of the AKR1B1 gene and cortical cataract in the presence of hyperglycemia.

Methods

In the second cross section of the Blue Mountains Eye Study (BMES), 3508 participants (2334 at 5-year follow-up and 1174 newly recruited participants) were examined during 1997 to 2000. Cataract was graded from lens photographs using the Wisconsin Cataract Grading System. Fasting blood glucose (FBG) was measured. Continuous imputed dosages of minor alleles of 17 AKR1B1 single nucleotide polymorphisms (SNPs) were assessed for associations with prevalent cortical cataract. Gene-environment interactions between SNPs and FBG were examined. Odds ratios (OR) and 95% confidence intervals (CI) for prevalent cortical cataract were estimated using logistic regression adjusting for age, sex, smoking, hypertension, education, and myopia. A P value of 0.005 was considered statistically significant after correction for 10 independent tests. Replication of significant associations found in the BMES sample was conducted in the Singapore Epidemiology of Eye Diseases (SEED) study (n = 10,033).

Results

No polymorphism was associated with prevalent cortical cataract. A significant interaction was observed between rs9640883 and FBG (Pinteraction = 0.004), with increased cortical cataract prevalence associated with rs9640883 minor allele dosage in those with FBG >6.0 mM (strata-specific OR 1.72, 95% CI 1.09-2.72). No similar association was found in participants with normal FBG (OR 0.85, 95% CI 0.69-1.04). This interaction was not evident in the SEED study.

Conclusions

The identified interaction between rs9640883 and FBG in relation to cortical cataract was not replicated but may warrant further investigation.

Prevalence, Risk Factors, and Impact of Undiagnosed Visually Significant Cataract: The Singapore Epidemiology of Eye Diseases Study

Objective

To determine the prevalence, risk factors, and impact of undiagnosed visually significant cataract in an Asian population.

Methods

The Singapore Epidemiology of Eye Diseases is a population-based study where 8,697 adults of Malay, Indian, and Chinese ethnicities aged > 40 years were invited for an eye examination, including lens photograph, to establish cataract diagnosis. Visually significant cataract was defined by Wisconsin Cataract Grading System and a best-corrected visual acuity <20/40 with cataract as the primary cause of vision impairment. Participants were deemed ‘undiagnosed’ if they had visually significant cataract and reported no prior physician diagnosis of cataract. Visual functioning (VF) was assessed with the VF-11 questionnaire validated using Rasch analysis.

Results

Among the 925 participants with visually significant cataract, 636 (68.8%) were unaware of their cataract status. Age-standardized prevalence varied according to ethnicity, with Malays having higher rates than Chinese and Indians. Factors independently associated with having undiagnosed visually significant cataract were: Malay ethnicity, lower educational attainment, in employment, and without a history of diabetes (all P<0.05). In those with undiagnosed visually significant cataract, half had bilateral visual impairment, which was significantly associated with 24.8% poorer visual functioning compared to those with unilateral visual impairment (P<0.001).

Conclusions

Two-thirds of Singaporean adults with visually significant cataract were previously undiagnosed. Half of these cases had bilateral visual impairment and substantially reduced quality of life. Public health strategies targeting elderly patients, such as regular screening for visual impairment and timely referral to ophthalmologists in order to prevent progression to bilateral visual impairment when visual function is compromised are warranted.

Associations Between Methylenetetrahydrofolate Reductase Polymorphisms, Serum Homocysteine Levels, and Incident Cortical Cataract

Importance

Methylenetetrahydrofolate reductase (MTHFR) polymorphisms have been shown to influence homocysteine levels; homocysteine has been implicated as a cataractogenic stressor.

Objective

To investigate the associations of MTHFR polymorphisms and serum homocysteine levels with incident cortical cataract in an older population.

Design, Setting, and Participants

From 1992 to 1994, a population-based cohort study, the Blue Mountains Eye Study, was conducted with 3654 residents (82.4% of eligible participants) of the Blue Mountains region aged 49 years and older. At the second (1997-1999, 5-year follow-up) and third (2002-2004, 10-year follow-up) surveys, 2334 (75.8% of survivors) and 1952 (76.7% of survivors) were examined, respectively. For this report, the second survey serves as baseline when homocysteine levels were assessed, and 5-year incidence of cataract refers to incidence estimated from the second to the third survey. After excluding participants with no follow-up data or DNA or who had previous cortical cataract or cataract surgery, 757 participants were included in gene and environment analyses. This current project on associations with cataract was designed initially March 19, 2013, and completed April 14, 2014. Cataract was assessed using the Wisconsin Cataract Grading system. Two MTHFR polymorphisms, C677T (rs1801133) and A1298C (rs1801131), were included. Serum homocysteine levels were assessed following standard methods.

Main Outcomes and Measures

Logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals for incident cortical cataract, after adjusting for age, sex, smoking status, hypertension, diabetes, education, and myopia. Path analysis was performed to explore a possible pathway of MTHFR polymorphisms via homocysteine levels to cortical cataract.

Results

The mean (SD) age of the 1726 participants in the Blue Mountains Eye Study 2 cohort with normal homocysteine levels was 68.3 (8.1) years and 73.2 (8.5) years for those with elevated homocysteine levels. Both the C677T polymorphism (CT/TT vs CC: OR = 1.50; 95% CI = 1.01-2.23) and elevated homocysteine levels (>15 µmol/L: OR = 2.24; 95% CI = 1.38-3.63) were independently associated with increased risk of cortical cataract. Path analysis showed that the genetic effect on cortical cataract was partially mediated via homocysteine levels. Combined CT/TT genotypes and elevated homocysteine levels were associated with a 3-fold risk of cortical cataract (OR = 3.74; 95% CI = 1.79-7.80). The synergy index of both exposures was 1.34 (95% CI = 0.44-4.01).

Conclusions and Relevance

MTHFR polymorphism and elevated homocysteine levels contributed separately and jointly to increased risk of cortical cataract. If these findings are confirmed, homocysteine levels may be a therapeutic target to reduce risk of cortical cataract in persons carrying genetic risk.

Ancestry, Socioeconomic Status, and Age-Related Cataract in Asians: The Singapore Epidemiology of Eye Diseases Study

PURPOSE

To determine the prevalence of age-related cataract and its ancestral and socioeconomic risk factors in a multi-ethnic Asian population.

DESIGN

Population-based, cross-sectional study.

PARTICIPANTS

A total of 10 033 adults (3353 Chinese, 3280 Malays, and 3400 Indians) aged >40 years in the Singapore Epidemiology of Eye Diseases Study.

METHODS

Study participants were invited for a structured interview and received a standardized comprehensive eye examination. Digital lens photographs were taken from eyes of each participant and graded for nuclear, cortical, and posterior subcapsular (PSC) cataract, following the Wisconsin Cataract Grading System. Prevalence data were compared with the Blue Mountains Eye Study (BMES) in Australia. Information on medical and lifestyle factors was collected using questionnaires and blood samples. To increase the precision of racial definition, genetic ancestry was derived from genome-wide single nucleotide polymorphism markers using principal component analysis. Regression models were used to investigate the association of cataract with socioeconomic factors (education and income) and genetic ancestry.

MAIN OUTCOME MEASURES

Age-related cataract.

RESULTS

A total of 8750 participants (94.0%) had gradable lens photographs. The age-standardized prevalence of cataract surgery in Chinese (16.0%), Malays (10.6%), and Indians (20.2%) was higher than in white subjects (4.1%). We found the age-standardized cataract prevalence in Chinese (30.4%), Malays (37.8%), and Indians (33.1%) was higher than in whites (18.5%). Cataract was 1.5 to 2 times more common in Asians and began 10 years earlier than in white subjects. Malays had significantly higher age-standardized prevalence of nuclear, cortical, and PSC cataract than Chinese (P<0.001). The severity of nuclear, cortical, and PSC cataract was significantly correlated with genetic ancestry in our South East Asian population. Less education and lower income were associated with cataract for Chinese and Indians but not Malays. The presence of visual impairment associated with cataract was higher in people aged ≥60 years and Malays.

CONCLUSIONS

We showed that people of different Asian ethnicities had a higher prevalence and earlier age of onset of cataract than Europeans. People of Malay ancestry have a greater severity for all cataract subtypes than people of Chinese ancestry. Education and income were associated with cataract for certain Asian subgroups.

Serum homocysteine, vitamin B12, and folate, and the prevalence and incidence of posterior subcapsular cataract

PURPOSE

We assessed associations between serum levels of homocysteine, vitamin B12, and folate, and the prevalence and 5-year incidence of posterior subcapsular cataract (PSC) in Blue Mountains Eye Study participants.

METHODS

We examined 3508 participants aged 49+ years during 1997 to 2000, including 2334 (75.1% of survivors) original and 1174 (85.2% of those eligible) newly recruited subjects. Five years later (2002-2004), 1952 (76.6% of survivors) original participants were re-examined. Detailed examinations, including lens photographs and fasting blood tests, were conducted at both visits. Logistic regression models estimated odds ratios (OR) and 95% confidence intervals (CI) after multivariable adjustment.

RESULTS

In this population, those with PSC were older, less likely to have higher education, and more likely to have diabetes and myopia. The PSC prevalence was 5.7% (150/2644). Higher levels of homocysteine (per SD; OR, 1.17; 95% CI, 1.00-1.37) and lower levels of folate (per SD; OR, 1.24; 95% CI, 0.99-1.56) were associated with prevalent PSC. There was significant interaction (P < 0.05) between vitamin B12 and homocysteine; for B12 ≥125 pmol/L, 28% higher PSC prevalence was associated with homocysteine (per SD; OR, 1.28; 95% CI, 1.09-1.52); however, for B12 <125 pmol/L, nonsignificant lower PSC prevalence was associated with homocysteine (per SD; OR, 0.16; 95% CI, 0.02-1.57). The 5-year PSC incidence was 5.7% (n = 59/1030) with no significant associations with homocysteine, B12, and folate.

CONCLUSIONS

Higher serum homocysteine level was associated with PSC prevalence in this population. Vitamin B12 status seemed to modify this association. Lack of longitudinal association could have resulted from insufficient study power.

Visual impairment corrected via cataract surgery and 5-year survival in a prospective cohort

PURPOSE

To compare mortality risk between cataract surgical patients with corrected and persistent visual impairment.

DESIGN

Cohort study.

METHODS

A total of 1864 consecutive patients, aged ≥64 years, undergoing phacoemulsification surgery at Westmead Hospital were followed annually for 5 years postoperatively. Visual impairment status in the surgical eye was categorized as none (presenting visual acuity [VA], ≥20/40), mild (VA <20/40-20/60), or moderate-severe (VA <20/60). All-cause mortality was obtained from the Australian National Death Index.

RESULTS

Of 901 patients with moderate-severe visual impairment before surgery, 60.4% (n = 544), 15.5% (n = 140), and 24.1% (n = 217) had no, mild, or moderate-severe visual impairment in the surgical eye, respectively, 1 month postoperatively. Age-standardized 5-year mortality rates were nonsignificantly lower in patients with either mild (24.7%, 95% confidence interval [CI] 16.5%-32.9%) or no visual impairment (24.1%, 95% CI 19.9%-28.4%) post surgery compared to that in patients whose moderate-severe visual impairment persisted (30.6%, 95% CI 23.3%-37.9%). After adjusting for age, sex, smoking, body mass index, and individual comorbid conditions, such as hypertension, diabetes, angina, myocardial infarction, stroke, transient ischemic attack, and kidney disease, patients with no visual impairment 1 month postoperatively had a lower mortality risk (HR 0.73, 95% CI 0.52-1.01) compared to those with persistent moderate-severe visual impairment after surgery. This finding was significant (HR 0.71, 95% CI 0.51-0.99) after additional adjustment for number of medications taken (continuous variable) and number (≥3 vs <3) of comorbid conditions.

CONCLUSION

Correcting moderate-severe visual impairment in older patients with phacoemulsification surgery was associated with a lower mortality risk, compared to surgical patients whose visual impairment persisted postoperatively.

Incidence and progression of reticular drusen in age-related macular degeneration: findings from an older Australian cohort

PURPOSE

To assess the 15-year incidence and progression of reticular drusen and associations of this lesion with age-related macular degeneration (AMD) risk factors.

DESIGN

Population-based cohort.

PARTICIPANTS

Blue Mountains Eye Study participants (n = 3654) 49 years of age and older attended baseline examinations; of these, 75.8%, 76.7%, and 56.1% of survivors attended 5-year, 10-year, and 15-year follow-up examinations, respectively.

METHODS

Color retinal photographs were obtained and comprehensive questionnaires were administered at each visit, and DNA samples were genotyped. Fundus autofluorescence images were not available. Reticular drusen identified from photographs were confirmed with side-by-side grading using the Wisconsin AMD grading protocol. Incidence was assessed using Kaplan-Meier product limit survival methods, controlling for competing risk of death. Associations between smoking, fish consumption, serum lipids, systemic and dietary factors, the CFH single nucleotide polymorphism (SNP) rs1061170 and ARMS2 SNP rs10490924, and the 15-year incidence of reticular drusen were analyzed in discrete logistic regression models. Generalized estimating equation models were used to analyze eye-specific relationships between these risk factors and 5-year progression from reticular drusen to late AMD.

MAIN OUTCOME MEASURES

Incidence and progression of reticular drusen.

RESULTS

The 15-year cumulative incidence of reticular drusen was 4.0% (n = 95). Increasing age (per decade increase; odds ratio [OR], 3.4; 95% confidence interval [CI], 2.6-4.4), female sex (OR, 2.0; 95% CI, 1.3-3.2), and presence of risk alleles of CFH-rs1061170 (OR, 1.8; 95% CI, 1.3-2.4) or ARMS2-rs10490924 (OR, 3.0; 95% CI, 2.1-4.4) were associated with higher reticular drusen incidence. Current smoking at baseline predicted higher reticular drusen incidence (OR 2.1, 95% CI 1.0-4.5) after adjusting for age, sex, CFH-rs1061170 and ARMS2-rs10490924 polymorphisms. Of 118 eyes with reticular drusen, 40 (33.9%) developed late AMD over 5 years. A higher proportion of eyes with reticular drusen located outside versus within the macular area progressed to late AMD (50.0% vs. 37.8%). Dietary lutein-zeaxanthin intake was associated with decreased likelihood of progression from reticular drusen to late AMD (adjusted OR, 0.5; 95% CI, 0.3-1.0).

CONCLUSIONS

Known AMD risk factors were associated with greater long-term risk of reticular drusen. Neither total area nor central location of reticular drusen predicted 5-year progression to late AMD. Increased consumption of lutein-zeaxanthin predicted a lower risk of progression.

Myopia, axial length, and age-related cataract: the Singapore Malay eye study

PURPOSE

To describe the associations of myopia and axial length (AL) with age-related cataract in an Asian population in Singapore.

METHODS

A population-based cross-sectional study that examined 3280 (78.7% response) adults of Malay ethnicity aged 40 to 80 years. Refractive error was determined by subjective refraction and AL was measured using the Zeiss IOL-Master. Digital slit lamp and retroillumination lens photographs were taken and graded for age-related nuclear, cortical, and posterior subcapsular (PSC) cataract following the Wisconsin system.

RESULTS

After excluding eyes with prior refractive or cataract surgery, 5474 eyes with gradable lens photographs were analyzed. In multivariate analyses adjusting for age, sex, body mass index, systolic blood pressure, glycosylated hemoglobin, smoking status, and education, myopia (spherical equivalent less than -0.5 diopter [D]) was associated with an increased prevalence of nuclear (OR: 4.99, 95% CI: 3.72–6.69) and PSC cataract (OR: 1.34, 95% CI: 1.30-1.39) but not with cortical cataract (OR: 0.85, 95% CI: 0.68-1.08) compared with emmetropia. Per-millimeter increase in AL was not associated with any of the three cataract subtypes. When myopia was defined as spherical equivalent of less than -5.0 D to -6.0 D, the OR of myopia for PSC cataract increased dramatically.

CONCLUSIONS

Our study shows that myopia, but not AL, was associated with nuclear cataract, supporting the concept of index myopia with aging. Myopia, especially high myopia, may predispose to PSC cataract formation. Clinically, ophthalmologists should be aware that risk of PSC cataract appears to vary by refractive status.

The impact of bilateral or unilateral cataract surgery on visual functioning: when does second eye cataract surgery benefit patients?

AIM

To examine the impact of bilateral or unilateral cataract surgery on visual functioning.

METHODS

The Singapore Malay Eye Study is a population-based study of 3280 Singapore Malay patients aged 40-80 years, of which 3225 had data available for inclusion. Cataracts were graded from digital lens photographs according to the Wisconsin scale. Study subjects were categorised as having: bilateral cataract surgery performed; unilateral cataract surgery performed with minimal cataract in the fellow eye; unilateral cataract surgery performed with significant cataract in the fellow eye; and bilateral cataract. Visual functioning was assessed using the modified VF-9 scale culturally adapted for Singaporean individuals, validated by Rasch analysis. The overall Rasch-modified vision-specific functioning score was compared across the four groups after adjusting for confounders such as age, gender, ocular and systemic comorbidities.

RESULTS

Persons with bilateral cataract had poorer visual functioning than those who had bilateral cataract surgery (mean visual functioning scores 3.38 vs 3.11, respectively, p=0.029). When compared with bilateral surgery, visual functioning improvements among patients with unilateral cataract surgery depended on the status of the fellow eye, with improvements only seen if the fellow eye had significant cataract (mean visual functioning scores 2.81 vs 3.25, p=0.019) or poor visual acuity (mean visual functioning scores 2.78 vs 3.25, p=0.018) after adjusting for confounders.

CONCLUSIONS

Bilateral cataract surgery was associated with greater visual functioning over unilateral cataract surgery when the fellow eye had a significant cataract or poor presenting visual acuity, supporting the current practice of second eye surgery depending on the fellow eye's cataract status and visual acuity.

A review and meta-analysis of the association between C-reactive protein and age-related macular degeneration

Age-related macular degeneration (AMD) is the leading cause of blindness in people over 60 in western countries. Inflammatory markers have been implicated in the development and progression of AMD. C-reactive protein (CRP) is an inflammatory marker known to be associated with cardiovascular disease, and a link between AMD and CRP has been suggested. In this systematic review we summarize the currently available evidence from clinic-based and population-based studies investigating this association. A meta-analysis of evidence from eleven studies (41,690 study participants) shows that high serum levels (>3 mg/L) of CRP are associated with a two-fold likelihood of late onset AMD, compared to low levels (<1mg/L). Sub-group meta-analyses show a higher association in studies using ophthalmoscopic examination, compared to those using photographic grading (pooled odds ratio 3.83 vs 1.36), to determine AMD status, or in clinic-based samples compared to population-based studies.

A computer-aided diagnosis system of nuclear cataract

Cataracts are the leading cause of blindness worldwide, and nuclear cataract is the most common form of cataract. An algorithm for automatic diagnosis of nuclear cataract is investigated in this paper. Nuclear cataract is graded according to the severity of opacity using slit lamp lens images. Anatomical structure in the lens image is detected using a modified active shape model. On the basis of the anatomical landmark, local features are extracted according to clinical grading protocol. Support vector machine regression is employed for grade prediction. This is the first time that the nucleus region can be detected automatically in slit lamp images. The system is validated using clinical images and clinical ground truth on >5000 images. The success rate of structure detection is 95% and the average grading difference is 0.36 on a 5.0 scale. The automatic diagnosis system can improve the grading objectivity and potentially be used in clinics and population studies to save the workload of ophthalmologists.

Use of inhaled and oral corticosteroids and the long-term risk of cataract

OBJECTIVE

Longitudinal associations between inhaled and oral corticosteroid use and 10-year incident cataract were examined.

DESIGN

Population-based cohort study.

PARTICIPANTS

The Blue Mountains Eye Study examined 3654 Australians aged 49 years or older (1992-1994); 2335 were re-examined after 5 years and 1952 were re-examined after 10 years (75.1%, 75.6% of survivors, respectively).

METHODS

Questionnaires were used to assess inhaled and oral corticosteroid use at baseline. Past users were participants who had used these medications for at least 1 month in the past but were not using them at baseline. Current users were those who were using these medications at baseline and had been doing so for at least 1 month. Ever users combined past and current users.

MAIN OUTCOME MEASURES

Lens photographs were obtained at each examination and graded for nuclear, cortical, and posterior subcapsular (PSC) cataracts following the Wisconsin Cataract Grading System. Participants without a specific subtype of cataract in either eye at baseline were considered to be at risk of that type of cataract developing over the 10-year follow-up. Incidence of each cataract subtype in this report refers to person-specific, first-eye incidence.

RESULTS

At baseline, 103 participants were current and 120 past users of inhaled corticosteroids, and 31 were current and 147 were past users of oral corticosteroids. Current users had a greater risk of developing PSC cataract after adjustment for age and gender (inhaled: odds ratio [OR] 2.50, 95% confidence interval [CI] 1.33-4.69; oral: OR 4.11; 95% CI 1.67-10.08) and nuclear cataract (inhaled: OR 2.04, 95% CI 1.21-3.43; oral: OR 3.45, 95% CI 1.26-9.43) but not cortical cataract. Interaction between inhaled and oral corticosteroid use was significant for PSC (P = 0.01) and nuclear (P = 0.02) cataract incidence. In subgroup analyses, only individuals who used both inhaled and oral steroids were at increased risk of PSC cataract (after adjusting for age, sex, smoking, hypertension, diabetes, and education levels; OR 4.76, 95% CI 2.59-8.74), comparing ever users of both with users of neither.

CONCLUSIONS

High long-term risks of PSC and nuclear cataract development were found for users of combined inhaled and oral corticosteroids.

Antioxidant nutrient intake and the long-term incidence of age-related cataract: the Blue Mountains Eye Study

BACKGROUND

Oxidative stress has been implicated in cataractogenesis. Long-term intake of antioxidants may offer protection against cataract.

OBJECTIVE

We investigated relations between antioxidant nutrient intakes measured at baseline and the 10-y incidence of age-related cataract.

DESIGN

During 1992-1994, 3654 persons aged >or=49 y attended baseline examinations of the Blue Mountains Eye Study (82.4% response). Of these persons, 2464 (67.4%) participants were followed >or=1 time after the baseline examinations (at either 5 or 10 y). At each examination, lens photography was performed and questionnaires were administered, including a 145-item semiquantitative food-frequency questionnaire. Antioxidants, including beta-carotene, zinc, and vitamins A, C, and E, were assessed. Cataract was assessed at each examination from lens photographs with the use of the Wisconsin Cataract Grading System. Nuclear cataract was defined for opacity greater than standard 3. Cortical cataract was defined as cortical opacity >or= 5% of the total lens area, and posterior subcapsular (PSC) cataract was defined as the presence of any such opacity.

RESULTS

Participants with the highest quintile of total intake (diet + supplements) of vitamin C had a reduced risk of incident nuclear cataract [adjusted odds ratio (OR): 0.55; 95% CI: 0.36, 0.86]. An above-median intake of combined antioxidants (vitamins C and E, beta-carotene, and zinc) was associated with a reduced risk of incident nuclear cataract (OR: 0.51; 95% CI: 0.34, 0.76). Antioxidant intake was not associated with incident cortical or PSC cataract.

CONCLUSION

Higher intakes of vitamin C or the combined intake of antioxidants had long-term protective associations against development of nuclear cataract in this older population.

Ten-year incidence of age-related cataract and cataract surgery in an older Australian population. The Blue Mountains Eye Study

PURPOSE

To estimate the 10-year incidence of cataract and cataract surgery in an older Australian population.

DESIGN

Prospective population-based study.

PARTICIPANTS

Persons at least 49 years old living in 2 postcode areas west of Sydney, Australia.

METHODS

Eye examinations were performed at baseline and at 5- and 10-year follow-up visits. Lens photographs were taken and graded by masked graders using the Wisconsin Cataract Grading System.

MAIN OUTCOME MEASURES

Incidences of nuclear cataract, cortical cataract, posterior subcapsular cataract (PSC), and cataract surgery.

RESULTS

Ten-year person-specific incidences were 36.0% for nuclear cataract, 28.0% for cortical cataract, 9.1% for PSC, and 17.8% for cataract surgery. Corresponding rates were 31.7%, 24.4%, 8.2%, and 14.4%, respectively, in men and 39.3%, 30.8%, 9.8%, and 20.1%, respectively, in women. The incidence for each type of cataract and cataract surgery was positively associated with age (P<0.0001). Women had a significantly higher incidence than men for nuclear cataract (P = 0.04), cortical cataract (P = 0.007), any cataract (P = 0.0006), and cataract surgery (P = 0.03) after adjusting for age. There was no significant gender difference for PSC. The mean age at cataract surgery was 75.8 years, and there was no significant gender difference (P = 0.9). Among persons who developed any cataract, 22% had more than one type and 1.3% had all 3 types present. Nuclear cataract and PSC were significantly associated with visual impairment (visual acuity worse than 20/40).

CONCLUSION

Age- and gender-specific cataract incidences in this study were similar to those reported from the U.S. Beaver Dam Eye Study. In this study, 72% of the participants were affected by cataract or had had cataract surgery over the 10-year follow-up period.

Circulating Inflammatory Markers and Hemostatic Factors in Age-Related Maculopathy: A Population-Based Case–Control Study

PURPOSE

To examine the relationship between circulating inflammatory markers, hemostatic factors, and age-related maculopathy (ARM).

METHODS

A population-based, cross-sectional case-control study drawn from the Blue Mountains Eye Study included 159 early and 38 late ARM cases, and 433 controls matched for age, gender, and smoking. ARM lesions were assessed from retinal photographs according to the Wisconsin ARM grading system. Circulating inflammatory markers (high-sensitivity C-reactive protein [hsCRP], intercellular adhesion molecule [ICAM]-1, and interleukin [IL]-6), white cell count (WCC), and hemostatic factors (fibrinogen, homocysteine, plasminogen activator inhibitor [PAI]-1 and von Willebrand factor [vWF]) were assessed. Age, gender, current smoking, body mass index, hypertension, history of stroke, and cardiovascular events were adjusted for. Adjusted mean levels of each marker were compared between persons with early ARM, those with late ARM, and control subjects, and are presented as probabilities. Adjusted associations with ARM were examined continuously (per SD), and are presented as odds ratios (ORs) and 95% confidence intervals (CIs). Summarizing z scores for inflammation and hemostatic dysfunction were calculated.

RESULTS

Increased PAI-1 level was associated with both early (OR 1.2, 95% CI 1.0-1.4 per SD increase) and late ARM (OR 1.3, 95% CI 0.9-1.9 per SD increase). Elevated ICAM-1 level was marginally associated with late ARM (OR 1.3, 95% CI 1.0-1.7 per SD increase). No other significant associations were found between the remaining inflammatory or hemostatic markers and either early or late ARM. Summarized z scores for inflammatory or hemostatic markers also did not suggest any associations.

CONCLUSIONS

There was no consistent pattern of association found between ARM and circulating inflammatory markers or hemostatic factors in this population-based case-control study.

Prevalence and Associations of Enhanced Retinal Arteriolar Light Reflex

PURPOSE

To assess the prevalence, associated risk factors and prognosis (mortality) of the enhanced retinal arteriolar light reflex sign in an older Australian population.

DESIGN

Population-based cross-sectional study.

PARTICIPANTS

Three thousand six hundred fifty-four participants (82.4% response) ages >/=49 years from Australia's Blue Mountains region.

METHODS

Retinal photographs of participants were graded for presence and severity of the enhanced arteriolar light reflex sign by comparison with standard photographs. Associations with systemic factors (subject-specific) and ocular variables (eye-specific) were assessed by logistic regression. Mortality data were obtained using the Australian National Death Index. Hazard ratios were calculated using Cox regression.

MAIN OUTCOME MEASURES

Prevalence of enhanced arteriolar light reflex and associations with demographic variables (age, gender), blood pressure, blood parameters, health risk behaviors, cataract, retinal vessel wall signs, retinopathy, and 10-year incident mortality.

RESULTS

The enhanced arteriolar light reflex sign was found in 1053 participants (31.7%, including 28.8% graded as mild and 2.9% as marked). Prevalence decreased with age (36.0%, 37.7%, 28.0%, and 18.8% for age groups < 60, 60-69, 70-79 and > or = 80 years, respectively, P(trend)<0.0001); odds ratio (OR) 0.78; and 95% confidence interval (CI) 0.72 to 0.85 per decade. Persons with cataract were less likely to have mildly enhanced light reflex (OR, 0.74; CI 0.64-0.87). After multivariate adjustment, mildly enhanced light reflex was significantly associated with serum glucose (OR 1.11 per SD increase), total cholesterol (OR 1.11), low-density lipoprotein (OR 1.55), triglycerides (OR 1.11), platelets (OR 0.89), and body mass index (OR 1.12). Markedly enhanced light reflex was significantly associated with mean arterial blood pressure (OR 1.24), heavy alcohol consumption (OR 2.66, > or = 40 grams alcohol per day), and serum glucose (OR 1.16). Strong associations were demonstrated between presence of mildly enhanced light reflex and either arteriovenous nicking (OR 3.12) or retinopathy (OR 1.96). There was no association between mildly or markedly enhanced light reflex and either all-cause or vascular mortality.

CONCLUSIONS

In this older population, the enhanced retinal arteriolar light reflex sign was a relatively common finding. Although some associations of this sign with vascular risk factors were found, only a marked level of enhanced light reflex was correlated with elevated blood pressure, but not with poor survival.

Comparison of age-specific cataract prevalence in two population-based surveys 6 years apart

Background

In this study, we aimed to compare age-specific cortical, nuclear and posterior subcapsular (PSC) cataract prevalence in two surveys 6 years apart.

Methods

The Blue Mountains Eye Study examined 3654 participants (82.4% of those eligible) in cross-section I (1992–4) and 3509 participants (75.1% of survivors and 85.2% of newly eligible) in cross-section II (1997–2000, 66.5% overlap with cross-section I). Cataract was assessed from lens photographs following the Wisconsin Cataract Grading System. Cortical cataract was defined if cortical opacity comprised ≥ 5% of lens area. Nuclear cataract was defined if nuclear opacity ≥ Wisconsin standard 4. PSC was defined if any present. Any cataract was defined to include persons who had previous cataract surgery. Weighted kappa for inter-grader reliability was 0.82, 0.55 and 0.82 for cortical, nuclear and PSC cataract, respectively. We assessed age-specific prevalence using an interval of 5 years, so that participants within each age group were independent between the two surveys.

Results

Age and gender distributions were similar between the two populations. The age-specific prevalence of cortical (23.8% in 1^st, 23.7% in 2^nd) and PSC cataract (6.3%, 6.0%) was similar. The prevalence of nuclear cataract increased slightly from 18.7% to 23.9%. After age standardization, the similar prevalence of cortical (23.8%, 23.5%) and PSC cataract (6.3%, 5.9%), and the increased prevalence of nuclear cataract (18.7%, 24.2%) remained.

Conclusion

In two surveys of two population-based samples with similar age and gender distributions, we found a relatively stable cortical and PSC cataract prevalence over a 6-year period. The increased prevalence of nuclear cataract deserves further study.

Increase in cataract surgery prevalence from 1992-1994 to 1997-2000: analysis of two population cross-sections

BACKGROUND

Monitoring temporal changes in the demand for and provision of cataract surgery will be useful to health planners to meet resource needs and to estimate likely costs. The aim of this report was to compare the prevalence of cataract surgery in two population cross-sections of the Blue Mountains Eye Study (BMES), over an interval of around 6 years.

METHODS

Cross-section I (n = 3654; mean age 66.2 years) consisted of BMES I participants (1992-1994). Cross-section II (n = 3509; mean age 66.7 years) consisted of BMES II (5-year) participants (n = 2335) plus persons who had moved into the area and age category since BMES I (n = 1174), so were newly eligible (1997-2000). Cataract surgery history was collected during interview and confirmed at examination.

RESULTS

Over this 6-year period, there was a 32% increase in cataract surgery prevalence from 6.0% to 7.9%, or from 6.0% to 7.7% after age standardization. The increase in eye-specific prevalence was 43% (from 4.4% to 6.3%). It was more marked among persons aged 80+ years and for bilateral (from 12.3% to 19.9%) than unilateral surgery (from 10.7% to 11.8%). Best-corrected visual acuity (mean letters read correctly) after surgery (43 and 44 letters, respectively) was similar between cross-sections I and II.

CONCLUSIONS

The findings from a representative older population confirm that prevalent cataract surgery has increased substantially over the average 6-year period, from 1992-1994 to 1997-2000.

Retinal vessel wall signs and the 5 year incidence of age related maculopathy: the Blue Mountains Eye Study

AIMS

To assess whether retinal arteriolar wall changes (focal narrowing and arteriovenous nicking) are associated with an increased 5 year risk of age related maculopathy (ARM).

METHODS

The Blue Mountains Eye Study examined 3654 residents aged 49+ years living in a defined area, during 1992-4 (82.4% participation). After 5 years, 2335 surviving participants (75.1%) were re-examined during 1997-9. Retinal photographs were graded using the Wisconsin ARM grading system. Incident late (neovascular or atrophic) or early stage ARM was defined using a side by side grading method. Focal arteriolar narrowing (localised constricted arteriolar segments causing a sausage-like appearance), and arteriovenous (AV) nicking (constriction on both sides of the venule where crossed by an arteriole), were graded by comparison with standard photographs. All retinal vessels passing through a circumferential zone 0.5-1.0 disc diameters from the optic disc margin were measured from digitised images. Summarised estimates for central retinal arteriolar equivalent (CRAE) represent an average diameter of arterioles for that eye. Associations were assessed after adjusting for age, sex, smoking, mean arterial blood pressure, and other vascular risk factors.

RESULTS

Of 2314 baseline participants at risk of late stage ARM, either late stage lesion developed in 34 participants (1.5%). Of 2203 at risk of early stage ARM, this sign developed in 197 participants (8.9%). Focal arteriolar narrowing was present at baseline in at least one eye of 162 survivors (6.9%) and severe AV nicking was present in 187 people (8.1%). Over 5 years, 4.9% of subjects with and 1.2% of those without focal narrowing developed either late stage ARM lesion, age adjusted relative risk (RR) 2.3, 95% confidence interval (CI) 1.0 to 5.1, multivariate adjusted odds ratios (OR) 2.1 (95% CI 0.9 to 4.9). Similarly, 3.7% of subjects with and 1.3% of those without severe AV nicking developed late ARM lesions, age adjusted RR 2.1 (95% CI 0.9 to 5.1), multivariate adjusted OR 2.2 (95% CI 0.9 to 5.6). Corresponding age adjusted RR and multivariate adjusted OR for development of early stage ARM were 1.4 (95% CI 0.9 to 2.0) and 1.3 (95% CI 0.8 to 2.1) for focal arteriolar narrowing, and 1.6 (95% CI 1.0 to 2.3) and 1.8 (95% CI 1.1 to 2.9) for severe AV nicking, respectively. No associations between baseline CRAE and 5 year incident late or early stage ARM were found.

CONCLUSIONS

Although of borderline statistical significance, the consistent associations found in this study suggest that structural retinal arteriolar changes may either contribute to ARM progression or may share common pathological pathways with ARM.

Clinical applications of MR cholangiopancreatography

INTRODUCTION

We review the clinical applications of magnetic resonance cholangiopancreatography (MRCP) and illustrate them with clinical examples.

METHODS

A literature search was performed in MedLine using the keywords "MR cholangiopancreatography" and "MRCP". The relevant articles were reviewed. The radiology information system was searched for MRCP examinations performed from July 1999 to December 2002. A surgical database of resected cases was also referenced. Illustrative cases were chosen in consultation with surgical authors to show the clinical applications of MRCP.

RESULTS

MRCP is useful for assessment of choledocholithiasis, choledochal cysts, pancreas divisum, primary sclerosing cholangitis, pancreatitis and anatomical variation of the biliary ducts. When complemented with dynamic gadolinium-enhanced studies, MRCP is useful for assessment of ductal cholangiocarcinoma, pancreatic and gall bladder neoplasms. Illustrative cases are provided.

CONCLUSION

MRCP is an important modality for the clinical assessment of biliary and pancreatic disorders.

Breath-hold fast recovery fast spin echo versus conventional non-breath-hold fast spin echo T2-weighted MR imaging of focal liver lesions

INTRODUCTION

We compare the breath-hold fast recovery fast spin echo (BHFRFSE) T2-weighted and non-breath-hold fast spin echo (NBHFSE) T2-weighted sequences in image quality and lesion characterisation of focal liver lesions.

MATERIALS AND METHODS

Fat-suppressed T2-weighted magnetic resonance (MR) images obtained with the 2 sequences (BHFRFSE and NBHFSE) in 79 patients with 113 liver lesions were analysed retrospectively. The image quality and nature of the lesions were evaluated by 2 experienced radiologists.

RESULTS

Based on receiver operating characteristic curve analysis, lesion characterisation was comparable for both sequences. The image quality of BHFRFSE was significantly better than that of NBHFSE. The NBHFSE missed 4 malignant lesions while BHFRFSE missed 2 malignant lesions.

CONCLUSION

BHFRFSE performs similarly to NBHFSE in image quality and liver lesion characterisation.

Real-time interactive duplex MR measurements: application in neurovascular imaging

OBJECTIVE

Real-time interactive duplex MR imaging is a new phase-contrast MR imaging technique that enables the quantification and display of flow velocities in real time without the need for cardiac gating. We investigated the feasibility and reliability of the technique to assess hemodynamic information both in vitro and in vivo in the carotid arteries and in the venous sinuses.

SUBJECTS AND METHODS

Real-time interactive duplex MR measurements (TR/TE, 53/27; flip angle, 90 degrees; encoding velocity, 100 or 150 cm/sec) were performed in vitro with a steady-flow phantom and in 10 healthy volunteers in whom common and internal carotid artery velocities were measured. In eight volunteers, velocity measurements were also performed in the superior sagittal sinus during both normal breathing and hyperventilation. Time-velocity plots were analyzed qualitatively and quantitatively and compared with findings from conventional segmented k-space phase-contrast MR imaging and Doppler sonography.

RESULTS

Velocity determinations for real-time duplex MR and conventional phase-contrast MR imaging showed an in vitro correlation of 0.99 and an in vivo correlation of 0.83 (carotid arteries) and 0.76 (venous sinus). Velocity measurements in the carotid arteries with real-time MR imaging were significantly lower than those obtained with conventional phase-contrast MR (averaged, 7.8%; p = 0.003) or sonography (23.7%, p < 0.001), likely because of volume averaging. Small but significant velocity changes occurring in the venous sinus during hyperventilation were reliably identified with both MR techniques.

CONCLUSION

Real-time interactive duplex MR imaging can be effectively applied in neurovascular imaging to obtain hemodynamic information.